Minority group inner-city adolescents are at particular risk for acquiring HIV/STDs. Interventions based on Social Cognitive theories have been successful in reducing risk, but the magnitude of effects has been small. We suggest that a neglected, but potentially important, set of variables may hold promise for improving risk reduction efforts: adolescent relationship factors. We propose to study adolescent couples in sexual relationships in order to identify how relationship characteristics and couple variables influence condom use. We believe that use of condoms is influenced by three levels of variables. (1) Individual variables such as beliefs about condom outcome expectancies and efficacy expectancies, and intention to use condoms; (2) Couple variables such as congruence/discrepancy in intention to use condoms with that partner as well as dynamics within the couple, including communication, power, and ambiguity and (3) Relationship variables including the type of relationship (messing, boy/girlfriend or hubby/wifey) and relationship characteristics, love, trust, expectation of monogamy, actual monogamy, salience and future commitment. The study has quantitative and qualitative components. Quantitative: We will recruit 360 "index" adolescents at baseline aged 14-17, also recruit a sexual partner (total n=720) and interview all 720 five times at 3 month intervals over one year whether or not they remain together as a couple. Baseline data analysis will examine sexual behavior and condom use as a function of individual beliefs, couple variables, and relationship variables. Longitudinal data analysis will examine relationship formation and dissolution, and identify factors associated with changes in risky sexual behavior over time within couples and in new relationships. We estimate that 25% of couples will remain together for 12 months and will provide data at all five data points; about 30% will be together for 3, 6 or 9 months and provide 2-4 data points; and about 45% of subjects will have couple data at baseline only. Qualitative research will use ethnographic interviews, cognitive interviews, focus groups, and couple interviews to design and test multi-item measures of relationship type and characteristics; to identify and eliminate sources of error in measuring condom use; to contribute to our theoretical understanding of how couple and relationship variables influence condom use, and to plan future HIV/STD preventive interventions. [unreadable] [unreadable]